Stats
Actions
Tags
From grounded-ai-dev
Phase 2 of Grounded AI — CHALLENGE. Use after a Requirements Document exists and before any tests or code, to stress-test the specification for gaps, ambiguities, and edge cases. Trigger when grounded-specify completes or when the user asks to review or pressure-test requirements. Produces a Specification Review Report and an updated Requirements Document (v2).
How this skill is triggered — by the user, by Claude, or both
Slash command
/grounded-ai-dev:grounded-challengeThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Test the specification before testing the code. This phase often uncovers the most valuable
Test the specification before testing the code. This phase often uncovers the most valuable
insights of the engagement — before any development cost is incurred. Save the output as
02-spec-review.md.
Ask and answer, in writing:
grounded-security-review for the threat-surface pass and grounded-ai-guardrails if
the system has generative features.Present v2 and request written client approval. No code, no tests until Gate A passes.
npx claudepluginhub pmj7671/grounded-ai-dev --plugin grounded-ai-devProvides CDSS development patterns for drug interaction checking, dose validation, clinical scoring (NEWS2, qSOFA), and alert classification integrated into EMR workflows.